|
Physics-Based Battery Insight
Artiphy infers model parameters, internal states and degradation behaviour from ordinary current, voltage and temperature measurements.
Machine-learning-accelerated surrogate models deliver orders-of-magnitude speedups over classical electrochemical solvers, enabling inference workflows that were previously impractical.
Battery modelling sits at the heart of modern electrification. Physics-based models can describe the internal electrochemistry of a cell and predict performance, degradation and safety across operating conditions.
But in practice, these models are often too slow, difficult to parameterise and hard to keep accurate as cells age or move into real-world use.
Artiphy closes this gap. We combine electrochemical modelling, machine learning and inverse-problem methods to infer hidden battery states and parameters from measurable signals such as current, voltage and temperature.
Where the physics is simple enough, we solve it directly. Where full simulations are too slow for inference, we use reduced-order models and AI surrogates to accelerate the workflow while retaining physical interpretability.
Next-generation battery intelligence powered by scientific machine learning
Estimate model parameters and stoichiometric alignment from current, voltage and temperature data, with uncertainty and identifiability checks.
ML-accelerated models make repeated electrochemical simulations practical for optimisation, uncertainty quantification and real-time inference workflows.
Infer internal states, transport limitations, ageing mechanisms and safety-relevant indicators that cannot be measured directly.
Our platform provides GPU-accelerated surrogate modelling for electrochemical battery systems, with a primary focus on the Single Particle Model (SPM).
Core workflows include:
ARTIPhy may be able to help you if you need to: